intelisys 0.3.8

Intelligence/AI services for the Lifsys Enterprise

Intelisys

Intelisys is a powerful Python library that provides a unified interface for interacting with various AI models and services. It offers seamless integration with OpenAI, Anthropic, OpenRouter, and Groq, making it an essential tool for AI-powered applications.

New in Version 0.3.8

• Prepared for PyPI update
• Incremented version number to 0.3.8 across all relevant files
• Ensured consistency in version numbers across the package
• Maintained all improvements from previous versions
• Updated documentation to reflect the latest changes
• [Add any new features or improvements specific to version 0.3.8 here]

Installation

Install Intelisys using pip:

pip install intelisys

For the latest development version:

pip install git+https://github.com/lifsys/intelisys.git

Requirements

• Python 3.7 or higher
• A 1Password Connect server (for API key management)
• Environment variables:
  • OP_CONNECT_TOKEN: Your 1Password Connect token
  • OP_CONNECT_HOST: The URL of your 1Password Connect server

Note: The library requires a local 1Password Connect server for API key retrieval.

Key Features

• Multi-provider support (OpenAI, Anthropic, OpenRouter, Groq)
• Secure API key management with 1Password Connect
• Asynchronous and synchronous chat interfaces
• Template-based API calls for flexible prompts
• JSON mode support for structured responses
• Lazy loading of attributes for improved performance
• Comprehensive error handling and logging
• Retry mechanism for API calls

Quick Start

from intelisys import Intelisys

# Using Intelisys class
intelisys = Intelisys(name="MyAssistant", provider="openai", model="gpt-4")
response = intelisys.chat("Explain quantum computing").results()
print(response)

# Using JSON mode
intelisys_json = Intelisys(name="JSONAssistant", provider="openai", model="gpt-4", json_mode=True)
response = intelisys_json.chat("List 3 quantum computing concepts").results()
print(response)  # This will be a Python dictionary

Advanced Usage

from intelisys import Intelisys
import asyncio

# Template-based API call
intelisys = Intelisys(name="TemplateAssistant", provider="anthropic", model="claude-3-5-sonnet-20240620")
render_data = {"topic": "artificial intelligence"}
template = "Explain {{topic}} in simple terms."
response = intelisys.template_chat(render_data, template).results()
print(response)

# Asynchronous chat
async def async_chat():
    intelisys = Intelisys(name="AsyncAssistant", provider="anthropic", model="claude-3-5-sonnet-20240620")
    response = await intelisys.chat_async("What are the implications of AGI?")
    print(await response.results())

asyncio.run(async_chat())

# Using context manager for temporary template and persona changes
intelisys = Intelisys(name="ContextAssistant", provider="openai", model="gpt-4")
with intelisys.template_context(template="Summarize {{topic}} in one sentence.", persona="You are a concise summarizer."):
    response = intelisys.template_chat({"topic": "quantum entanglement"}).get_last_response()
    print(response)

# Using retry mechanism
intelisys = Intelisys(name="RetryAssistant", provider="openai", model="gpt-4", max_retry=5)
response = intelisys.chat("This might fail, but we'll retry").get_last_response()
print(response)

Supported Providers and Models

Intelisys supports a wide range of AI providers and models:

• OpenAI: Various GPT models including gpt-4
• Anthropic: Claude models including claude-3-5-sonnet-20240620
• OpenRouter: Access to multiple AI models through a single API
• Groq: Fast inference models

For a complete list of supported models, please refer to the DEFAULT_MODELS dictionary in the Intelisys class.

Error Handling

Intelisys now includes improved error handling and a retry mechanism for API calls. If an API call fails, the library will automatically retry the call up to the specified max_retry times (default is 10). This helps to handle temporary network issues or API rate limits.

JSON Parsing

For JSON responses, Intelisys now uses a more robust parsing method. If the standard json.loads() fails, it falls back to safe_json_loads() from the utilisys library, which can handle some common JSON parsing errors.

API Reference

For detailed information on available methods and their usage, please refer to the docstrings in the source code or our API documentation.

Contributing

We welcome contributions! Please see our Contributing Guidelines for more details.

License

This project is licensed under the MIT License - see the LICENSE file for details.

Changelog

For a detailed list of changes and version history, please refer to the CHANGELOG.md file.

About Lifsys, Inc

Lifsys, Inc is an innovative AI company dedicated to developing cutting-edge solutions for the future. Visit www.lifsys.com to learn more about our mission and projects.